🤗 Transformers: State-of-the-art Machine Learning for Pytorch, TensorFlow, and JAX.

DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective.

Ray is an AI compute engine. Ray consists of a core distributed runtime and a set of AI Libraries for accelerating ML workloads.

An LLM-powered knowledge curation system that researches a topic and generates a full-length report with citations.

Fully open reproduction of DeepSeek-R1

Universal LLM Deployment Engine with ML Compilation

[ICLR 2024] Efficient Streaming Language Models with Attention Sinks

verl: Volcano Engine Reinforcement Learning for LLMs

Training and serving large-scale neural networks with auto parallelization.

Line-by-line profiling for Python

PyTorch native quantization and sparsity for training and inference

Comprehensive and timely academic information on federated learning (papers, frameworks, datasets, tutorials, workshops)

计算机网络-自顶向下方法 习题/编程/实验答案

Implementation of Communication-Efficient Learning of Deep Networks from Decentralized Data

xDiT: A Scalable Inference Engine for Diffusion Transformers (DiTs) with Massive Parallelism

Rotary Transformer

Local models support for Microsoft's graphrag using ollama (llama3, mistral, gemma2 phi3)- LLM & Embedding extraction

My learning notes/codes for ML SYS.

MoBA: Mixture of Block Attention for Long-Context LLMs

Implements harmful/harmless refusal removal using pure HF Transformers

Deep Learning Energy Measurement and Optimization

Push-Button End-to-End Testing of Kubernetes Operators and Controllers

ShiftAddLLM: Accelerating Pretrained LLMs via Post-Training Multiplication-Less Reparameterization

Google TPU optimizations for transformers models

Tele-LLMs is a series of open-source large language models specialized in telecommunications